[mac] Init spellcheck service.

[mac] Init spellcheck service.

BUG=454992
Committed: https://crrev.com/4ca14cb4922b883fde5419b9b176ef58b5296bf2
Cr-Commit-Position: refs/heads/master@{#317179}

[please use gerrit instead, 2015-02-13 16:42:59]

rouslan@chromium.org changed reviewers:
+ groby@chromium.org

[please use gerrit instead, 2015-02-13 16:42:59]

This is not the final patch, but it fixes spellcheck on Mac already. I just
wanted to upload it so you can appreciate the weirdness of the situation...

[please use gerrit instead, 2015-02-13 23:32:23]

Patchset #1 (id:1) has been deleted

[please use gerrit instead, 2015-02-13 23:32:28]

Patchset #1 (id:20001) has been deleted

[please use gerrit instead, 2015-02-13 23:34:05]

Rachel, ptal.

In an ideal world, I would probably rearchitect many things in spellcheck. But
given that this is a regression that needs to merge to dev (I think), I'd like
to keep the patch small.

[groby-ooo-7-16, 2015-02-18 17:51:55]

Where does Views instantiate the factory? Think you can find a way to guarantee
a common initialization path for all platforms?

[please use gerrit instead, 2015-02-18 20:30:55]

Views initializes spellcheck service in the following locations, which are not
used on Mac.

SpellCheckMessageFilter::OnSpellCheckerRequestDictionary - Mac does not have a
dictionary.
SpellCheckMessageFilter::OnCallSpellingService - Mac has its own copy of
spellcheck service.

I would unify more Mac and Views code an ideal world, but that's out of scope
for fixing this particular regression.

[groby-ooo-7-16, 2015-02-19 00:21:00]

Alright, alright... LGTM

[please use gerrit instead, 2015-02-19 00:26:46]

The CQ bit was checked by rouslan@chromium.org

[commit-bot: I haz the power, 2015-02-19 00:28:22]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/928473002/40001

[commit-bot: I haz the power, 2015-02-19 01:50:31]

The CQ bit was unchecked by commit-bot@chromium.org

[commit-bot: I haz the power, 2015-02-19 01:50:32]

Try jobs failed on following builders:
  mac_chromium_rel_ng on tryserver.chromium.mac (JOB_FAILED,
http://build.chromium.org/p/tryserver.chromium.mac/builders/mac_chromium_rel_...)

[please use gerrit instead, 2015-02-19 02:39:52]

The CQ bit was checked by rouslan@chromium.org

[commit-bot: I haz the power, 2015-02-19 02:40:30]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/928473002/40001

[commit-bot: I haz the power, 2015-02-19 03:54:38]

The CQ bit was unchecked by commit-bot@chromium.org

[commit-bot: I haz the power, 2015-02-19 03:54:39]

Try jobs failed on following builders:
  mac_chromium_rel_ng on tryserver.chromium.mac (JOB_FAILED,
http://build.chromium.org/p/tryserver.chromium.mac/builders/mac_chromium_rel_...)

[please use gerrit instead, 2015-02-19 15:31:12]

Failure is reliable. Looking into it.

[please use gerrit instead, 2015-02-19 23:07:43]

New patchsets have been uploaded after l-g-t-m from groby@chromium.org

[please use gerrit instead, 2015-02-19 23:10:32]

Moved service initialization from constructor to
SpellCheckMessageFilterMac::OnRequestTextCheck, because it needs to happen on
the UI thread. This was the reason for the unit test failures.

[please use gerrit instead, 2015-02-19 23:21:53]

The CQ bit was checked by rouslan@chromium.org

[commit-bot: I haz the power, 2015-02-19 23:23:37]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/928473002/60001

[groby-ooo-7-16, 2015-02-19 23:41:15]

I'm not happy with doing that in RequestTextCheck - that's a cost for every
check we do request. Grumble. _Especially_ since you only do this to satisfy a
unit test which doesn't _have_ a UI thread. 

The proper fix would be to send an init message from cocoa_spelling_engine, just
like we do for hunspell_engine. (Except that's yet one more path that's
different for Cocoa. So, no)

Sigh.

[commit-bot: I haz the power, 2015-02-20 00:07:31]

Committed patchset #2 (id:60001)

[commit-bot: I haz the power, 2015-02-20 00:07:59]

Patchset 2 (id:??) landed as
https://crrev.com/4ca14cb4922b883fde5419b9b176ef58b5296bf2
Cr-Commit-Position: refs/heads/master@{#317179}

[please use gerrit instead, 2015-02-20 15:22:21]

On 2015/02/19 23:41:15, groby wrote:
> I'm not happy with doing that in RequestTextCheck - that's a cost for every
> check we do request. Grumble. _Especially_ since you only do this to satisfy a
> unit test which doesn't _have_ a UI thread. 
> 
> The proper fix would be to send an init message from cocoa_spelling_engine,
just
> like we do for hunspell_engine. (Except that's yet one more path that's
> different for Cocoa. So, no)
> 
> Sigh.

Let me know if you'd like me to revert and look for a better solution.